Ann. rheum. Dis. (1971), 30, 375 Ann Rheum Dis: first published as 10.1136/ard.30.4.375 on 1 July 1971. Downloaded from

Palindromic Clinical and immunological studies

M. H. WILLIAMS,* P. J. H. S. SHELDON,t G. TORRIGIANI, V. EISEN, AND S. MATTINGLY Departments of Research, Rheumatology and Physical Medicine, and Immunology, Middlesex Hospital Medical School and Middlesex Hospital, London

Hench and Rosenberg (1944) described 34 patients tions were less frequent, and the finger pads were not with recurring episodes of and periarthritis involved. The duration of attacks was longer, usually lasting less than a week and named by them intervals between attacks tended to be shorter, and 'palindromio rheumatism'. The features they empha- there were often symptoms between attacks. The sized may be summarized as follows: erythrooyte sedimentation rate was increased, and (1) Recurrent attacks ofjoint pain and swelling at x rays andjoint biopsy showed changes characteristic variable and irregular intervals lasting a few of . Nevertheless, some authors hours or a few days. have expressed the view that palindromic rheuma- (2) Any joint affected but especially fingers, tism is a variant of episodic rheumatoid arthritis or a wrists, shoulders and knees. stage in its development (Ansell and Bywaters, 1959; (3) Para-artioular attacks and transient nodules. Robinson, 1963; Mattingly, 1966). (4) Good health: normal tests and x-rays. copyright. (5) Good prognosis-no effective treatment. Present investigations Several case reports have appeared since then, though few series of patients have been followed up In an attempt to differentiate palindromic rheuma- (Ward and Okihiro, 1959; Rotes Querol and Lience, tism from rheumatoid arthritis on an immunological 1959; Dames and Zuckner, 1961; Ansell and basis, leucocyte migration inhibition and anti-IgG Bywaters, 1959). levels were estimated. The migration of leucocytes The main differential diagnosis is from crystal in the presence of Mycoplasma fermentans antigen http://ard.bmj.com/ synovitis which may be due either to sodium urate or has been shown to be inhibited in two-thirds of to calcium pyrophosphate. In these conditions, cases of seropositive rheumatoid arthritis (Williams, attacks are usually monarticular, occurring in Brostoff, and Roitt, 1970) and anti-IgG levels were favoured joints and frequently lasting for about a found to be raised in cases of seronegative rheuma- week. In those due to sodium urate, erosions may be toid arthritis (Torrigiani, Roitt, Lloyd, and Corbett, detected, there is hyperuricaemia, and crystals of 1970). These studies were therefore carried out in the monosodium urate may be detected in tophi and group of patients diagnosed clinically as having on September 26, 2021 by guest. Protected synovial fluid. In those due to calcium pyrophos- palindromic rheumatism and compared with results phate, calcification ofarticular cartilage may be seen, obtained in patients with rheumatoid arthritis. and the diagnosis is supported by the demonstration Several cases featuring pain, swelling, and dis- of calcium pyrophosphate crystals in synovial fluid; coloration around joints were ascribed by Cohen hyperuricaemia may or may not be associated. (1911) to angioneurotic oedema. These he believed Intermittent hydrarthrosis is a condition dis- to be 'frequently mistaken for or rheumatism'. tinguished by the regularity of attacks, which Patients with hereditary angioneurotic oedema have nearly always affect one or both knees, though been shown to lack a component of the complement other joints may rarely be involved (Mattingly, system-C'1 esterase inhibitor (Donaldson and 1957). Rosen, 1966). Webb (1970) described a patient in The differentiation of palindromic rheumatism whom a diagnosis of palindromic rheumatism was from episodic rheumatoid arthritis is not easy, though entertained but who was subsequently shown to have Hench (1947) mentioned that the latter tended to hereditary angioneurotic oedema, confirmed by the occur in favoured joints, para-articular manifesta- absenoe of C'1 esterase inhibitor in the patient's

* Present address: Nuffield Unit of Medical Genetics, University of Liverpool. t Reprint requests: Dr. P. J. H. S. Sheldon, Department of Rheumatology and Physical Medicine, Middlesex Hospital, London, WIP 9PG. Accepted for publication February 12, 1971.

D 376 Annals of the Rheumatic Diseases Ann Rheum Dis: first published as 10.1136/ard.30.4.375 on 1 July 1971. Downloaded from

serum. In view of the possibility that the episodic Houses) was used and a 0 5 stock solution was prepared nature of palindromic rheumatism may be regarded in 2-ethoxy-ethanol. as a joint manifestation analogous to the skin and A human serum fraction rich in C'1 esterase was suboutaneous tissue phenomena seen in hereditary separated and activated by the method of Nelson (1965). Preparations containing 60 to 110 C'1 esterase units per angioneurotic oedema, C'1 esterase inhibitor was ml. were obtained (the hydrolysis of ATEe by one C'1 estimated in our series of patients. esterase unit releases 33-3 n-mole of H+ per minute: Levy and Lepow, 1959). Serum levels of C'1 esterase inhibitor were measured Patients studied by a modification of the method of Levy and Lepow (1959). Ten ,ul. of serum were incubated with 4-5 units of 35 patients (13 male and 22 female) with a diagnosis of C'1 esterase in a Radiometer titration vessel for 3 min. palindromic rheumatism were investigated in the Depart- Saline and ATEe were then added and the hydrolysis of ment of Rheumatology and Physical Medicine of the 0-025M ATEe, in a total volume of 2 ml. followed by Middlesex Hospital. The duration of history varied from continuous automatic titration with 5mM NaOH at 9 mths to 23 yrs (mean 7-4 yrs). The diagnostic criteria 37°C. and pH 7 2. The protein concentration in the adopted were: system was less than 1 mg./ml.; no buffer was added. (1) Attacks of pain and swelling in any joint occurring With this method, values of 18-35 inhibitor units/ml. at irregular intervals. were found in normal sera (one inhibitor unit inhibits ten (2) The subsequent return of the state of that joint to C'1 esterase units, i.e. the release of 333 n-mole of its condition prior to the attack. H+/min.: Levy and Lepow, 1959). (3) Duration of symptoms of a few hours to a few days. (4) The absence when the diagnosis was first made, of chronic , subcutaneous nodules or Results erosions on x-ray. (5) Serum uric acid less than 6- 5 mg/ml. CLINICAL FINDINGS (6) The absence of calcification in articular cartilage. The ratio of females:males in the 35 patients was 1 * 7:1 (compared with 3:1 for rheumatoid arthritis- Miall, Ball, and Kellgren, 1958). Twelve patients Material and methods gave a history of purely monarticular attacks, threecopyright. of only polyarticular attacks, and sixteen of both. ANTIGLOBULINS Included as polyarticular attacks were those cases in Sera from patients were screened for the presence of antiglobulins () using the latex test which subsiding in one joint was (Hyland, rheumatoid arthritic test latex-globulin followed by the onset of inflammation in another. reagent). The quantitative estimation of these antibodies This could occur before the first joint had com- was carried out as described by Torrigiani and others pletely returned to its original state before the attack. (1970), using horse immunoglobulin as antigen. The In the remainder the patient was either unsure, or the http://ard.bmj.com/ normal range was taken as 15 to 35 ,g./ml. question was omitted at the time of interview. The minimum duration of patients' attacks was 2 hrs, the MYCOPLASMA ANTIGENS longest being about 1 week. These were prepared as described previously (Williams The joints involved are depicted in Table I which and others, 1970). Briefly, the method involved growing shows that almost any joint may be affected. The Mycoplasma fermentans (P.G. 18) in standard liquid upper limb was more commonly involved than the medium. Organisms were harvested from 5-litre cultures

lower, the spine rarely. on September 26, 2021 by guest. Protected which had been incubated for 36 hrs. Cells were washed four times in 0 *2M phosphate buffer pH 7 * 2. Membranes Table I Joints involved at any time during course of were collected by centrifugation of cells which had been disease in 35 patients with palindromic rheumatism disrupted by hypotonic shock and sonication (Williams and Taylor-Robinson, 1967). The protein concentration of the membrane suspension was adjusted to approx- Joint No. ofpatients imately 10 mg./ml. Temporo-mandibular 7 Shoulder 23 LEUCOCYTE MIGRATION Elbow 21 The method followed was that of Bendixen and S0borg Wrist 29 the that leucocytes Carpus 25 (1969). This exploits phenomenon Metacarpophalangeal 25 from sensitized individuals do not migrate from a capillary Proximal interphalangeal 24 tube in the presence of antigen as well as they do in its Terminal interphalangeal 13 absence. Inhibition was considered to be significant when Spine 4 the area of migration was less than 80 per cent. of that Hip 14 seen in the control chambers lacking antigen. Knee 26 Ankle 21 Foot 18 C'1 -ESTERASE Toe 19 N-acetyl-L-tyrosine ethyl ester (ATEe; British Drug Palindromic rheumatism 377 Ann Rheum Dis: first published as 10.1136/ard.30.4.375 on 1 July 1971. Downloaded from

The age at onset of symptoms is depicted in Fig. 1; Nineteen of the 35 patients had a low-grade most cases commenced in the fifth decade, and there arthropathy at the time of interview, and two were was no difference between the sexes. seen during an acute attack of palindromic rheuma- 20 tism. In the latter, the joint was extremely tender male with swelling and there was redness of the skin overlying thejoint. These patients are distinguishable 1 _Ir temale a)'A 15 by an asterisk in Table II. Two males and one female 0 had rheumatoid nodules at the time of examination. u 0 Sixteen cases had no arthropathy 3 mths to 12 yrs 10- after onset of symptoms (mean 5 95 yrs). Of these, -o E ten were female. z 5, INVESTIGATIONS Antiglobulins 0 The anti-IgG levels were found to be raised in all but 1-10 11-20 21-30 31-40 41-50 51-60 61+ one of the patients (Fig. 2, overleaf). Elevation of Age at onset (yrs ) anti-IgG levels may be a reflection of a disorder of FIG. 1 Age at onset ofpalindromic rheumatism immune complexes. Similar values have been found

Table II Clinical and laboratory findings in 35 cases ofpalindromic rheumatism

Patient Sex Age at Duration of Arthropathy Leucocyte Anti-IgG C'esterase Latex at No. onset disease at time of migrationt levels inhibitor time of (yrs) when seen review (per cent.) (g./ml.) (units/ml.)§ review 11 (yrs)

1 F 27 4 59 93 18-3 weak + copyright. 2 41 5 80 61 14-4 + 3 39 5 + 50 74 14-8 weak + 4 59 4 64 74 18-6 weak + 5 38 10 53 74 19-6 + 6* 49 12 50 61 32 5 + 7 49 4k 68 94 36 + 8 39 S + 85 72 42 weak + 9 49 6 83 58 26-2 weak + 10 47 11 75 105 33 + http://ard.bmj.com/ 11 42 7 + 52 95 25-3 + 12 46 2 + 58 65 26-4 + 13 46 7 65 53 15X9 + 14 45 9 + 81 61 24-3 + 15 56 2k 78 175 25 + 16 35 18 + 93 76 30 + 17 45 6 mths 84 81 22 3 weak + 18 40 8 78 95 22 3 weak + 19 49 12 87 61 27 6 on September 26, 2021 by guest. Protected 20 29 16 68 150 33 + 21 48 8k 88 174 27 1 + 22 63 3 mths 35 220 18-3 + + 23* M 53 9 mths 43 47 20X2 weak + 24 57 10 mths 37 48 23-1 weak + 25 57 4 + 63 93 17-1 + 26 43 4 60 61 24-1 27 31 4k 57 69 14 8 + 28 39 4k 63 88 21 3 29 32 6 79 57 18*6 + 30 55 6 93 88 19-2 + 31 43 7 56 72 30 3 wek+ 32 44 8 40 72 14-7 + 33 46 11 87 88 20 8 34 29 4 66 15 21 3 + 35 22 23 24 74 16'8 +

* Patient seen during acute attack of palindromic rheumatism. t Normal > 80 per cent. § Normal 18 to 35 units/ml. S Normal 15 to 35 i&g/m1. 11 Weak + corresponds to 1:40 or laes. 378 Annals of the Rheumatic Diseases Ann Rheum Dis: first published as 10.1136/ard.30.4.375 on 1 July 1971. Downloaded from in Still's disease (Torrigiani, Ansell, Chown, and Of the eight males, two remained seropositive and Roitt, 1969), seronegative rheumatoid arthritis one seronegative, and three converted from sero- (Torrigiani and Roitt, 1967), and procainamide- negative to seropositive, and two from seronegative induced systemic lupus erythematosus (Sheldon and to weakly positive. Williams, 1970). 220 Leucocyte migration inhibition In the presence of Mycoplasma fermentans antigen 200 this test was unequivocally positive in 25 of the 35 patients with palindromic rheumatism (i.e. 71 per 180 cent.). These results are shown in Fig. 4 and are 160 compared with the previously published results obtained with comparable groups of seropositive __JE 140 00 rheumatoid arthritics and normal healthy subjects (Williams and others, 1970). tQ 120 1201 v 0 80 4" 0 .< 0 60 100. 0* 40 0 a so 20 0*@ 0 = 80. 0 normal palindromic rheumatoid acs S. normal polindromic rheumatoid 0. * subjects rheumatoids arthritis E 4, FIG. 2 Anti-IgG levels in 35 cases of palindromic c rheumatism with nineteen cases rheumatoid copyright. compared of 0 60 arthritis and sixteen controls 0:0 0@ 0@0 Rheumatoidfactor In 23 patients (15 females and 8 males), the original 0 0 40 result of rheumatoid factor estimation was known. 0 The subsequent course of these patients is depicted S in Fig. 3. http://ard.bmj.com/ 0 20 I 3 normal palindromic rheumatoid subjects rheumatoids arthritis F,G. 4 Leucocyte migration results in 35 cases ofpalin- S. dromic rheumatism compared with 42 cases ofrheumatoid O* latex and SCAT ne ative arthritis and 21 controls * - latex or SCAT weak3 positive C *m latex or S.C.AT positive i. 0 - S on September 26, 2021 by guest. Protected latex titre :1:40 or - 0 S.CATI titre ?1:32 1 * C'1 esterase The patients were not deficient in C'1 esterase O_ inhibitor, although fairly low levels were found in 0 ~ ~ 0. seven out of 35 sera (Table II). Ci0-~ 35 Discussion The clinical impression gained when examining the was that over half of were 1960'b1 '62 '63 '64 '65 '66 'b7 '68 '69 197O patients just (19 35) Year indistinguishable from patients with low-grade FIG. 3 Results of rheumatoid factor test in 23 patents rheumatoid arthritis. However, sixteen patients had with palindromic rheumatism no arthropathy when seen, 3 mths to 12 yrs (mean 5 9 yrs) after the commencement of symptoms. Of Of the fifteen females, three remained seropositive, these only two remained seronegative. and six converted from seronegative to seropositive, Serum levels of C'1 esterase inhibitor showed no one from seropositive to weakly positive, and five significant aberration in any of the patients. from seronegative to weakly positive. Leuoocyte migration inhibition in the presence of Palindromic rheumatism 379 Ann Rheum Dis: first published as 10.1136/ard.30.4.375 on 1 July 1971. Downloaded from

Mycoplasma fermentans antigen was similar in the palindromic rheumatism were reviewed olinic- groups irrespective of the presence of arthropathy ally and immunologically at a mean of 7 *4 yrs and resembled the results obtained in rheumatoid after the onset of symptoms. arthritis. This also applied to the IgG antiglobulin (2) On clinical grounds nineteen appeared to have estimations, and suggests that, in these two respects, developed a low-grade rheumatoid arthritis. the patients with palindromic attacks were indis- Sixteen, however, had no arthropathy. tinguishable from those with rheumatoid arthritis. (3) Serum anti-IgG levels were raised in 34 cases, No correlation was found between antiglobulin the results being similar to those found in levels and the degree ofleucocyte migration inhibition seronegative rheumatoid arthritis. in the presence of Mycoplasma fermentans antigen (4) The leucocyte migration inhibition test in the (r =003). The elevated level of anti-IgG may arise presence of Mycoplasma fermentans antigen in two ways: was positive in 25 cases, the results being similar to those found in rheumatoid (1) Nonspecific immunoglobulin is bound to the arthritis. membrane of Mycoplasma fermentans, is thereby (5) Serum levels of C'esterase inhibitor were denatured and becomes antigenic (Williams and normal in 28 cases, the remainder being others, 1970). slightly, though not significantly, lower than (2) Specific antibody to Mycoplasma fermentans under- normal. goes conformational change, resulting in the forma- (6) These observations provide further evidence tion of a new antigen, thus stimulating the production for palindromic rheumatism being a pro- of antiglobulin. dromal or atypical presentation of rheumatoid arthritis, with an essentially benign prognosis. The inhibition of leucocyte migration may indicate a state of cell mediated (delayed type) hypersensitivity We wish to thank Dr. A. C. Boyle and Dr. Mary Corbett (Williams and others, 1970) to Mycoplasma fermen- for allowing us to carry out these studies on patients tans antigen. under their care. We also thank Prof. C. A. Keele for his help and advice and Miss Norma Banks and Mr. R.

Perry for their excellent technical assistance. The work copyright. Summary was supported by grants from the Arthritis and Rheuma- tism Council, the Wates Foundation, and the Wcllcome (1) 35 patients diagnosed initially as having Trust. http://ard.bmj.com/

References ANSELL, B. M., AND BYwATERS, E. G. L. (1959) Ann. rheum. Dis., 18, 331 (Palindromic rheumatism).

BENDIXEN, G., AND S0BORG, M. (1969) Dan. med. Bull., 16, 1 (A leucoCYte migration technique for in vitro on September 26, 2021 by guest. Protected detection of cellular-rapid delayed type-hypersensitivity in man). COHEN, S. SOLIs (1911) Trans. Coil. Phys. Philadelphia, 3rd ser., 33, 309 (Certain angioneurotic manifestations in and around joints frequently mistaken for gout or rheumatism). DAMEs, R., AND ZUCKNER, J. (1961) Arch. interamer. Rheum., 4, 18 (Palindromic rheumatism). DONALDSON, V. H., AND ROsEN, F. S. (1966) Pediatrics, 37, 1017 (Hereditary angio-neurotic oedema-a clinica survey). HENcH, P. S. (1947) Arizona Med., 4, 62 (Arizona medical problems consultation and case analysis). - AND ROsENBERG, E. F. (1944) Arch. intern. Med., 73, 293 (Palindromic rheumatism, a new oft recurring disease of joints). LEvy, L. R., AND LEPOw, I. H. (1959) Proc. Soc. exp. Biol. (N. Y.), 101, 608 (Assay and properties of serum inhibitors of C'1 esterase). MATmrNGLY, S. (1957) Brit. med. J., 1, 139 (Intermittent hydrarthrosis). - (1966) Ann. rheum. Dis., 25, 307 (Palindromic rheumatism). MIALL, W. E., BALL, J., AND KELLGREN, J. H. (1958) Ibid., 17, 263 (Prevalence of rheumatoid arthritis in urban and rural populations in South Wales). NELsoN, R. A. (1965) 'The role of complement in immune phenomena', in 'The Inflammatory Process', ed. B. W. Zweifach, L. Grant, and R. T. McCluskey, p. 819 (1965) Academic Press, New York. ROBINSON, W. D. (1963) In 'Cecil-Loeb Textbook of Medicine', 11th ed., ed. P. B. Beeson and W. McDermott, p. 1484. Saunders, Philadelphia. 380 Annals of the Rheumatic Diseases Ann Rheum Dis: first published as 10.1136/ard.30.4.375 on 1 July 1971. Downloaded from

RoTrs QUEROL, J., AND IJENCE, E. (1959) Arch. interamer. Rheum., 2, 491 (Rheumatismos intermitentes). SHELDON, P. J. H. S., AND WILLIAMS, W. R. (1970) Ann. rheum. Dis., 29, 236 (Procainamide-induced S.L.E.). TORRIGIANI, G., ANSELL, B. M., CHOWN, E. E. A., AND Rorrr, I. M. (1969) Ann. rheum. Dis., 28, 424 (Raised IgG antiglobulin factors in Still's disease). AND Rorrr, I. M. (1967) Ibid., 26, 334 (Antiglobulin factors in sera from patients with rheumatoid arthritis and normal subjects). -, , LLOYD, K. N., AND CORBETr, M. (1970) Lancet, 1, 14 (Elevated IgG antiglobulin in patients with sero-negative rheumatoid arthritis). WARD, L. E., AND OKHIRO, M. M. (1959) Quoted by R. W. Lamont-Havers (1960) In 'Arthritis and Allied Conditions', ed. J. L. Hollander, p. 708. Lea and Febiger, Philadelphia. WEBB, F. W. S. (1970) Proc. roy. Soc. Med., 63, 281 (Hereditary angio-oedema). WILLIAMS, M. H., BROSTOFF, J., AND Roirr, I. M. (1970) Lancet, 2, 277 (Possible role of Mycoplasma fermentans in pathogenesis of rheumatoid arthritis). - AND TAYLOR-ROBINSON, D. C. (1967) Nature (Lond.), 215, 973 (Antigenicity of mycoplasma membranes). copyright. http://ard.bmj.com/ on September 26, 2021 by guest. Protected